First results from DAMA/LIBRA

Transcription

1 First results from DAMA/LIBRA R. Cerulli INFN-LNGS DSU 2008 BUE, Cairo, Egypt Jun 1-5,

2 Roma2,Roma1,LNGS,IHEP/Beijing DAMA/NaI DAMA/LIBRA DAMA/LXe DAMA/R&D low bckg DAMA/Ge for sampling meas.

3 Relic DM particles from primordial Universe etc etc SUSY (as neutralino or sneutrino In various scenarios) the sneutrino in the Smith and Weiner scenario sterile ν electron interacting dark matter a heavy ν of the 4-th family even a suitable particle not yet foreseen by theories axion-like (light pseudoscalar and scalar candidate) self-interacting dark matter mirror dark matter etc Kaluza-Klein particles (LKK) heavy exotic canditates, as 4th family atoms,... Elementary Black holes, Planckian objects, Daemons & (& invisible axions, ν s) Right halo model and parameters? Composition? DM multicomponent also in the particle part? Right related nuclear and particle physics? Non thermalized components? Caustics? clumpiness?

4 Some direct detection processes: and more Scatterings on nuclei detection of nuclear recoil energy DMp N DMp Ionization: Ge, Si Bolometer: TeO 2, Ge, CaWO 4,... Scintillation: NaI(Tl), LXe,CaF 2 (Eu), Inelastic Dark Matter:W W + N W* + N W has Two mass states χ+, χ- withδ mass splitting Kinematical constraint for the inelastic scattering of χ- on a nucleus 1 2 µv2 δ v v thr = 2δ µ Excitation of bound electrons in scatterings on nuclei detection of recoil nuclei + e.m. radiation Interaction only on atomic electrons detection of e.m. radiation DMp e -... even WIMPs e.g. sterile ν Conversion of particle into e.m. radiation detection of γ, X-rays, e - Interaction of ligth DMp (LDM) on e - or nucleus with production of a lighter particle detection of electron/nucleus recoil energy a X-ray e - γ e.g. signals from these candidates are completely lost in experiments based on rejection procedures of the e.m. component of their rate

5 The annual modulation: a model independent signature for the investigation of Dark Matter particles component in the galactic halo With the present technology, the annual modulation is the main model independent signature for the DM signal. Although the modulation effect is expected to be relatively small a suitable large-mass, low-radioactive set-up with an efficient control of the running conditions would point out its presence. Drukier, Freese, Spergel PRD86 Freese et al. PRD88 30 km/s December v sun ~ 232 km/s (Sun velocity in the halo) v orb = 30 km/s (Earth velocity around the Sun) γ = π/3 ω = 2π/T T = 1 year t 0 = 2 nd June (when v is maximum) June 30 km/s ~ 232 km/s 60 Requirements of the annual modulation S k v (t) = v sun + v orb cosγcos[ω(t-t 0 )] dr [ η( t)] = der S0, k + Sm, k cos[ ω( t t0)] de E k R Expected rate in given energy bin changes because the annual motion of the Earth around the Sun moving in the Galaxy 1) Modulated rate according cosine 2) In a definite low energy range 3) With a proper period (1 year) 4) With proper phase (about 2 June) 5) For single hit events in a multi-detector set-up 6) With modulation amplitude in the region of maximal sensitivity must be <7% for usually adopted halo distributions, but it can be larger in case of some possible scenarios To mimic this signature, spurious effects and side reactions must not only - obviously - be able to account for the whole observed modulation amplitude, but also to satisfy contemporaneously all the requirements

6 Competitiveness of NaI(Tl Tl) ) set-up Well known technology High duty cycle Large mass possible Ecological clean set-up; no safety problems Cheaper than every other considered technique Small underground space needed High radiopurity by selections, chem./phys. purifications, protocols reachable Well controlled operational condition feasible Routine calibrations feasible down to kev range in the same conditions as the production runs Neither re-purification procedures nor cooling down/warming up (reproducibility, stability,...) Absence of microphonic noise + effective noise rejection at threshold (τ of NaI(Tl) pulses hundreds ns, while τ of noise pulses tens ns) High light response ( ph.e./kev) Sensitive to SI, SD, SI&SD couplings and to other existing scenarios, on the contrary of many other proposed target-nuclei Sensitive to both high (by Iodine target) and low mass (by Na target) candidates Effective investigation of the annual modulation signature feasible in all the needed aspects Fragmented set-up etc. A low background NaI(Tl) also allows the study of several other rare processes such as: possible processes violating the Pauli exclusion principle, CNC processes in 23 Na and 127 I, electron stability, nucleon and dinucleon decay into invisible channels, neutral SIMP and nuclearites search, solar axion search,... High benefits/cost

7 The 100 kg NaI(Tl) ) set-up (DAMA/NaI NaI) 9 highly radiopure NaI(Tl) of 9,7 kg coupled throught 10 cm long Tetrasil-B light guides to 2 low background PMTs (specially developed). Detectors enclosed in a sealed copper box, continuously maintained in HP Nitrogen atmosphere, in slightly overpressure with respect to the external environment. A suitable low background hard shield against e.m. and neutron background was realized using very high radiopure Cu and Pb bricks, Cd foils and 10/40 cm polyethylene/paraffin; the hard shield is also sealed in a plexiglas box and maintained in the high purity HP Nitrogen atmosphere. Etc. Model independent evidence for the presence of DM particles in the galctic halo Performances: N.Cim.A112(1999) , EPJC18(2000)283, Riv.N.Cim.26 n. 1(2003)1-73, IJMPD13(2004)2127 Results on rare processes: Possible Pauli exclusion principle violation CNC processes Electron stability and non-paulian transitions in Iodine atoms (by L-shell) Search for solar axions Exotic Matter search Search for superdense nuclear matter Search for heavy clusters decays Results on DM particles: PLB408(1997)439 PRC60(1999) PLB460(1999)235 PLB515(2001)6 EPJdirect C14(2002)1 EPJA23(2005)7 EPJA24(2005)51 PSD PLB389(1996)757 Investigation on diurnal effect N.Cim.A112(1999)1541 Exotic Dark Matter search PRL83(1999)4918 Annual Modulation Signature PLB424(1998)195, PLB450(1999)448, PRD61(1999)023512, PLB480(2000)23, EPJC18(2000)283, PLB509(2001)197, EPJC23(2002)61, PRD66(2002)043503, Riv.N.Cim.26 n.1 (2003)1, IJMPD13(2004)2127, IJMPA21(2006)1445, EPJC47(2006)263, IJMPA22(2007)3155, EPJC53(2008)205, PRD77(2008) arxiv: , other works in progress... data taking completed on July 2002, last data release 2003; still producing results on corollary quests for the candidate particle and the possible astrophysical, nuclear and particle physics scenarios total exposure ( 7 annual cycles) 0.29 ton x yr

8 The model-independent evidence by DAMA/NaI well compatible with several candidates in several of the many astrophysical, nuclear and particle physics scenarios; other ones are open PRD69(2004) Neutralino as LSP in SUSY theories Various kinds of WIMP candidates with several different kind of interactions Pure SI, pure SD, mixed + Migdal effect +channeling, (from low to high mass) WIMP with preferred inelastic scattering a heavy ν of the 4-th family Pseudoscalar, scalar or mixed light bosons with axion-like interactions Mirror Dark Matter Light Dark Matter Dark Matter (including some scenarios for WIMP) electron-interacting Elementary Black holes such as the Daemons and more Sterile neutrino Self interacting Dark Matter heavy exotic canditates, as 4th family atoms,... Kaluza Klein particles Possible model dependent positive hints from indirect Available results from direct searches searches not in conflict with DAMA/NaI result using different target materials and approaches (but interpretation, evidence itself, derived mass and from searches for up-going muons and cross sections depend e.g. on bckg modeling, on do not give any robust conflict DM spatial velocity distribution in the galactic halo, etc.)

9 The new DAMA/LIBRA set-up ~250 kg NaI(Tl) (Large sodium Iodide Bulk for RAre processes) As a result of a second generation R&D for more radiopure NaI(Tl) by exploiting new chemical/physical radiopurification techniques (all operations involving crystals and PMTs - including photos - in HP Nitrogen atmosphere) installing DAMA/LIBRA detectors assembling a DAMA/ LIBRA detector detectors during installation; in the central and right up detectors the new shaped Cu shield surrounding light guides (acting also as optical windows) and PMTs was not yet applied filling the inner Cu box with further shield closing the Cu box housing the detectors view at end of detectors installation in the Cu box

10 Dismounting/Installing protocol (with Scuba system) All the materials selected for low radioactivity Multicomponent passive shield Three-level system to exclude Radon from the detectors Calibrations in the same running conditions as production runs Installation in air conditioning + huge heat capacity of shield Monitoring/alarm system; many parameters acquired with the prod data Pulse shape recorded by Waweform Analyzer TVS641A (2chs per detector), 1 Gsample/s, 8 bit, bandwidth 250 MHz Data collected from low energy up to MeV region, despite the hardware optimization was done for the low energy Polyethylene/ paraffin 25 x 9.7 kg NaI(Tl) in a 5x5 matrix two Suprasil-B light guides directly coupled to each bare crystal two PMTs working in coincidence at the single ph. el. threshold Installation Glove-box for calibration For details, radiopurity,performances, procedures, etc. see arxiv:0804:2738 to appear on NIMA Electronics + DAQ 1m concrete from GS rock

11 Some on residual contaminants in new NaI(Tl) detectors e α α/e pulse shape discrimination has practically 100% effectiveness in the MeV range The measured α yield in the new DAMA/LIBRA detectors ranges from 7 to some tens α/kg/kev Second generation R&D for new DAMA/LIBRA crystals: new selected powders, physical/chemical radiopurification, new selection of overall materials, new protocol for growing and handling live time = 570 h Th residual contamination From time-amplitude method. If 232 Th chain at equilibrium: it ranges from 0.5 ppt to 7.5 ppt 238 U residual contamination First estimate: considering the measured α and 232 Th activity, if 238 U chain at equilibrium 238 U contents in new detectors typically range from 0.7 to 10 ppt 238 U chain splitted into 5 subchains: 238 U 234 U 230 Th 226 Ra 210 Pb 206 Pb Thus, in this case: (2.1±0.1) ppt of 232 Th; (0.35 ±0.06) ppt for 238 U and: (15.8±1.6) µbq/kg for 234 U Th; (21.7±1.1) µbq/kg for 226 Ra; (24.2±1.6) µbq/kg for 210 Pb nat K residual contamination The analysis has given for the nat K content in the crystals values not exceeding about 20 ppb double coincidences 129 I and 210 Pb 129 I/ nat I for all the new detectors 210 Pb in the new detectors: (5 30) µbq/kg.... more on arxiv: No sizeable surface pollution by Radon daugthers, thanks to the new handling protocols

12 DAMA/LIBRA: calibrations at low energy Studied by using various external gamma sources ( 241 Am, 133 Ba) and internal X-rays or gamma s ( 40 K, 125 I, 129 I) The curves superimposed to the experimental data have been obtained by simulations Internal 40 K: 3.2 kev due to X-rays/Auger electrons (tagged by 1461 kev γ in an adiacent detector). Internal 125 I: 67.3 kev peak (EC from K shell kev γ) and composite peak at 40.4 kev (EC from L,M,.. shells kev γ). External 241 Am source: 59.5 kev γ peak and 30.4 kev composite peak. External 133 Ba source: 81.0 kev γ peak. Internal 129 I: 39.6 kev structure (39.6 kev γ + β spectrum). 3.2 kev 241 Am Internal 40 K Tagged by an adjacent detector 30.4 kev 59.5 kev Internal 125 I first months 40.4 kev 133 Ba 67.3 kev 81 kev Linearity Energy resolution 39.6 kev Internal 129 I σ LE E ( ± 0.035) 3 ( ) 10 = + ± E( kev ) Routine calibrations with 241 Am

13 DAMA/LIBRA: calibrations at high energy Cs Co 662 kev 1173 kev 1332 kev The data are taken on the full energy scale up to the MeV region by means QADC s 2505 kev Studied by using external sources of gamma rays (e.g. 137 Cs, 60 Co and 133 Ba) and gamma rays of 1461 kev due to 40 K decays in an adjacent detector, tagged by the 3.2 kev X-rays 133 Ba 40 K 81 kev 356 kev 1461 kev Linearity Energy resolution σ HE E ( 1.12 ± 0.06) 4 ( 17 23) 10 = + ± E( kev ) The signals (unlike low energy events) for high energy events are taken only from one PMT

14 Noise rejection near the energy threshold Typical pulse profiles of PMT noise and of scintillation event with the same area, just above the energy threshold of 2 kev PMT noise The different time characteristics of PMT noise (decay time of order of tens of ns) and of scintillation event (decay time about 240 ns) can be investigated building several variables Single-hit production data γ source Scintillation event 2-4 kev X 2 X 1 PMT noise 4-6 kev X 2 X 2 X 1 X 1 X 1 X 2 Scintillation pulses From the Waveform Analyser 2048 ns time window: Area (from 100 ns to 600 ns) X= 1 ; Area (from 0 ns to 600 ns) Area (from 0 ns to 50 ns) X= 2 Area (from 0 ns to 600 ns) The separation between noise and scintillation pulses is very good. Very clean samples of scintillation events selected by stringent acceptance windows. The related efficiencies evaluated by calibrations with 241 Am sources of suitable activity in the same experimental conditions and energy range as the production data (efficiency measurements performed each ~10 days; typically events per kev collected) This is the only procedure applied to the analysed data

15 Infos about DAMA/LIBRA data taking DAMA/LIBRA test runs: from March 2003 to September 2003 DAMA/LIBRA normal operation: from September 2003 to August 2004 High energy runs for TDs: September 2004 to allow internal α s identification (approximative exposure 5000 kg d) DAMA/LIBRA normal operation: from October 2004 arxiv: Data released here: four annual cycles: 0.53 ton yr calibrations: acquired 44 M events from sources acceptance window eff: acquired 2 M events/kev DAMA/NaI (7 years) + DAMA/LIBRA (4 years) Two remarks: total exposure: kg day = 0.82 ton yr One PMT problems after 6 months. Detector out of trigger since Sep (it will be put again in operation at the 2008 upgrading) Residual cosmogenic 125I presence in the first year in some detectors (this motivates the Sept as starting time) DAMA/LIBRA is continuously running

16 Cumulative low-energy distribution of the single-hit scintillation events Single-hit events = each detector has all the others as anticoincidence (Obviously differences among detectors are present depending e.g. on each specific level and location of residual contaminants, on the detector s location in the 5x5 matrix, etc.) DAMA/LIBRA (4 years) total exposure: 0.53 ton yr Efficiencies already accounted for About the energy threshold: The DAMA/LIBRA detectors have been calibrated down to the kev region. This assures a clear knowledge of the physical energy threshold of the experiment. It obviously profits of the relatively high number of available photoelectrons/kev (from 5.5 to 7.5). 3.2 kev, tagged by 1461 kev γ in an adjacent detector The two PMTs of each detector in DAMA/LIBRA work in coincidence with hardware threshold at single photoelectron level. Effective near-threshold-noise full rejection. The software energy threshold used by the experiment is 2 kev.

17 Experimental single-hit residuals rate vs time and energy Model-independent investigation of the annual modulation signature has been carried out by exploiting the time behaviour of the residual rates of the single-hit events in the lowest energy regions of the DAMA/LIBRA data. These residual rates are calculated from the measured rate of the single-hit events (obviously corrections for the overall efficiency and for the acquisition dead time are already applied) after subtracting the constant part: r ijk flat 30 km/s jk December jk r ijk is the rate in the considered i-th time interval for the j-th detector in the k-th energy bin June 30 km/s ~ 232 km/s 60 flat jk is the rate of the j-th detector in the k-th energy bin averaged over the cycles. The average is made on all the detectors (j index) and on all the energy bins (k index) The weighted mean of the residuals must obviously be zero over one cycle.

18 Model Independent Annual Modulation Result DAMA/NaI (7 years) + DAMA/LIBRA (4 years) Total exposure: kg day = 0.82 ton yr experimental single-hit residuals rate vs time and energy arxiv: Acos[ω(t-t 0 )] ; continuous lines: t 0 = d, T = 1.00 y 2-4 kev A=(0.0215±0.0026) cpd/kg/kev χ 2 /dof = 51.9/ σ C.L. Absence of modulation? No χ 2 /dof=117.7/67 P(A=0) = kev A=(0.0176±0.0020) cpd/kg/kev χ 2 /dof = 39.6/ σ C.L. Absence of modulation? No χ 2 /dof=116.1/67 P(A=0) = kev A=(0.0129±0.0016) cpd/kg/kev χ 2 /dof = 54.3/ σ C.L. Absence of modulation? No χ 2 /dof=116.4/67 P(A=0) = The data favor the presence of a modulated behavior with proper features at 8.2σ C.L.

19 Model-independent residual rate for single-hit events DAMA/NaI (7 years) + DAMA/LIBRA (4 years) total exposure: kg day = 0.82 ton yr Results of the fits keeping the parameters free: A (cpd/kg/kev) T= 2π/ω (yr) t 0 (day) C.L. DAMA/NaI (7 years) (2 4) kev ± ± ± σ (2 5) kev ± ± ± σ (2 6) kev ± ± ± σ DAMA/LIBRA (4 years) (2 4) kev ± ± ± σ (2 5) kev ± ± ± 9 6.9σ (2 6) kev ± ± ± σ DAMA/NaI + DAMA/LIBRA (2 4) kev ± ± ± 7 8.3σ (2 5) kev ± ± ± 7 8.9σ (2 6) kev ± ± ± 8 8.2σ

20 Modulation amplitudes, A, of single year measured in the 11 one-year experiments of DAMA (NaI + LIBRA) The difference in the (2 6) kev modulation amplitudes between DAMA/NaI and DAMA/LIBRA depends mainly on the rate in the (5 6) kev energy bin. The modulation amplitudes for the (2 6) kev energy interval, obtained when fixing exactly the period at 1 yr and the phase at days, are: (0.019 ± 0.003) cpd/kg/kev for DAMA/NaI (0.011 ± 0.002) cpd/kg/kev for DAMA/LIBRA. Thus, their difference: (0.008 ± 0.004) cpd/kg/kev is 2σ which corresponds to a modest, but non negligible probability. Moreover: The χ 2 test (χ 2 = 4.9, 3.3 and 8.0 over 10 d.o.f. for the three energy intervals, respectively) and the run test (lower tail probabilities of 74%, 61% and 11% for the three energy intervals, respectively) accept at 90% C.L. the hypothesis that the modulation amplitudes are normally fluctuating around their best fit values. Compatibility among the annual cycles

21 Clear annual modulation is evident in (2-6) kev while it is absence just above 6 kev Power spectrum of single-hit residuals Treatment of the experimental errors and time binning included here DAMA/NaI (7 years) total exposure: 0.29 ton yr 2-6 kev vs 6-14 kev DAMA/LIBRA (4 years) total exposure: 0.53 ton yr (according to Ap.J.263(1982)835;.263(1982)835; Ap.J.338(1989)277).338(1989)277) DAMA/NaI (7 years) + DAMA/LIBRA (4 years) total exposure: 0.82 ton yr 2-6 kev 6-14 kev 2-6 kev 6-14 kev 2-6 kev 6-14 kev Principal mode in the 2-6 kev region: DAMA/NaI DAMA/LIBRA DAMA/NaI+LIBRA d -1 1 y d -1 1 yr d -1 1 yr -1 + Not present in the 6-14 kev region (only aliasing peaks)

22 Can a hypothetical background modulation account for the observed effect? No Modulation above 6 kev A=(0.9±1.1) 10-3 cpd/kg/kev DAMA/LIBRA Mod. Ampl. (6-10 kev): ( ± ), -( ± ), -( ± ) and -( ± ) cpd/kg/kev for DAMA/LIBRA-1, DAMA/LIBRA-2, DAMA/LIBRA-3, DAMA/LIBRA-4; they can be considered statistically consistent with zero + In the same energy region where the effect is observed: no modulation of the multiple-hits events (see next slide) No modulation in the whole spectrum: R 90 percentage variations with respect to their mean values for single crystal in the DAMA/LIBRA-1,2,3,4 running periods Fitting the behaviour with time, adding a term modulated according period and phase expected for Dark Matter particles: studying integral rate at higher energy, R90 cumulative gaussian behaviour with σ 1%, fully accounted by statistical considerations Period Mod. Ampl. DAMA/LIBRA-1 -(0.05±0.19) cpd/kg DAMA/LIBRA-2 -(0.12±0.19) cpd/kg DAMA/LIBRA-3 -(0.13±0.18) cpd/kg DAMA/LIBRA-4 (0.15±0.17) cpd/kg consistent with zero + if a modulation present in the whole energy spectrum at the level found in the lowest energy region R 90 tens cpd/kg 100 σ far away σ 1% No modulation in the background: these results account for all sources of bckg (+ see later)

23 Multiple-hits events in the region of the signal - DAMA/LIBRA 1-4 Each detector has its own TDs read-out pulse profiles of multiple-hits events (multiplicity > 1) acquired (exposure: 0.53 ton yr). Initial time August, kev: A mult =-(0.0004±0.0008) cpd/kg/kev The same hardware and software procedures as the ones followed for single-hit events signals by Dark Matter particles do not belong to multiple-hits events, that is: 2 5 kev: A mult =-(0.0005±0.0007) cpd/kg/kev multiple-hits events = Dark Matter particles events switched off 2 6 kev: A mult =-(0.0004±0.0006) cpd/kg/kev Evidence of annual modulation with proper features as required by the DM annual modulation signature is present in the single-hit residuals, while it is absent in the multiple-hits residual rate. This result offers an additional strong support for the presence of Dark Matter particles in the galactic halo further excluding any side effect either from hardware or from software procedures or from background

24 Energy distribution of the modulation amplitudes, S m, for the total exposure [ ( t )] R( t) = S0 + Sm cos ω t heret=2π/ω=1 yr and t 0 = day 0 DAMA/NaI (7 years) + DAMA/LIBRA (4 years) total exposure: kg day = 0.82 ton yr E = 0.5 kev bins A clear modulation is present in the (2-6) kev energy interval, while S m values compatible with zero are present just above In fact, the S m values in the (6 20) kev energy interval have random fluctuations around zero with χ 2 equal to 24.4 for 28 degrees of freedom

25 Statistical distributions of the modulation amplitudes (S( m ) a) S m values for each detector, each annual cycle and each considered energy bin (here 0.25 kev) b) <S m > = mean values over the detectors and the annual cycles for each energy bin; σ = errors associated to each S m DAMA/LIBRA (4 years) total exposure: 0.53 ton yr Each panel refers to each detector separately; 64 entries = 16 energy bins in 2-6 kev energy interval 4 DAMA/LIBRA annual cycles 2-6 kev Standard deviations of the variable (S m - S m )/σ for the DAMA/LIBRA detectors r.m.s < r.m.s. < 1.16 Individual S m values follow a normal distribution since (S m -<S m >)/σ is distributed as a Gaussian with a unitary standard deviation (r.m.s.) S m statistically well distributed in all the detectors and annual cycles

26 Statistical analyses about modulation amplitudes (S( m ) x=(s m -<S m >)/σ, χ 2 =Σ x 2 χ 2 /d.o.f. values of S m distributions for each DAMA/LIBRA detector in the (2 6) kev energy interval for the four annual cycles. The line at χ 2 /d.o.f. = 1.31 corresponds to an upper tail probability of 5%. DAMA/LIBRA (4 years) total exposure: 0.53 ton yr Comparison with χ 2 distribution with 64 d.o.f. gives: χ 2 /d.o.f. = 8.1/7 The χ 2 /d.o.f. values range from 0.7 to 1.28 (64 d.o.f. = 16 energy bins 4 annual cycles) at 95% C.L. the observed annual modulation effect is well distributed in all the detectors. The mean value of the twenty-four points is 1.072, slightly larger than 1. Although this can be still ascribed to statistical fluctuations, let us ascribe it to a possible systematics. In this case, one would have an additional error of cpd/kg/kev, if quadratically combined, or cpd/kg/kev, if linearly combined, to the modulation amplitude measured in the (2 6) kev energy interval. This possible additional error ( 4.7% or 0.7%, respectively, of the DAMA/LIBRA modulation amplitude) can be considered as an upper limit of possible systematic effects

27 Is there a sinusoidal contribution in the signal? Phase day? [ ω( t t )] + Z sin ω( t t ) [ ] [ ( )] * = S + Y t R( t) = S ω t 0 + Sm cos 0 m 0 0 m cos For Dark Matter signals: Z m «S m Y m t * t 0 = 152.5d ω = 2π/T T = 1 year Slight differences from 2 nd June are expected in case of contributions from non thermalized DM components (as e.g. the SagDEG stream) E (kev) S m (cpd/kg/kev) Z m (cpd/kg/kev) Y m (cpd/kg/kev) t* (day) ± ± ± ± ± ± ±

28 The analysis at energies above 6 kev, the analysis of the multiple-hits events and the statistical considerations about S m already exclude any sizeable presence of systematical effects. Additional investigations on the stability parameters Modulation amplitudes obtained by fitting the time behaviours of main running parameters, acquired with the production data, when including a DM-like modulation Running conditions stable at a level better than 1% DAMA/LIBRA-1 DAMA/LIBRA-2 DAMA/LIBRA-3 DAMA/LIBRA-4 Temperature -( ± ) C ( ± ) C (0.001 ± 0.015) C ( ± ) C Flux N 2 (0.13 ± 0.22) l/h (0.10 ± 0.25) l/h -(0.07 ± 0.18) l/h -(0.05 ± 0.24) l/h Pressure (0.015 ± 0.030) mbar -(0.013 ± 0.025) mbar (0.022 ± 0.027) mbar ( ± ) mbar Radon -(0.029 ± 0.029) Bq/m 3 -(0.030 ± 0.027) Bq/m 3 (0.015 ± 0.029) Bq/m 3 -(0.052 ± 0.039) Bq/m 3 Hardware rate above single photoelectron -(0.20 ± 0.18) 10-2 Hz (0.09 ± 0.17) 10-2 Hz -(0.03 ± 0.20) 10-2 Hz (0.15 ± 0.15) 10-2 Hz All the measured amplitudes well compatible with zero +none can account for the observed effect (to mimic such signature, spurious effects and side reactions must not only be able to account for the whole observed modulation amplitude, but also simultaneously satisfy all the 6 requirements)

29 Example of Stability Parameters: DAMA/LIBRA-1 R Hj = hardware rate of j-th detector above single photoelectron Operating Temperature HP Nitrogen flux Radon external to the shield HP N 2 Pressure in the inner Cu box Running conditions stable at level < 1% Parameters distributions All amplitudes well compatible with zero + no effect can mimic the annual modulation

30 Temperature Detectors in Cu housings directly in contact with multi-ton shield huge heat capacity ( 10 6 cal/ 0 C) Experimental installation continuosly air conditioned (2 independent systems for redundancy) Operating T of the detectors continuously controlled σ=0.4% Amplitudes for annual modulation in the operating T of the detectors well compatible with zero DAMA/LIBRA-1 DAMA/LIBRA-2 DAMA/LIBRA-3 DAMA/LIBRA-4 T ( C) -( ± ) ( ± ) (0.001 ± 0.015) ( ± ) Distribution of the root mean square values of the operating T within periods with the same calibration factors (typically 7days): mean value 0.04 C Considering the slope of the light output -0.2%/ C: relative light output variation < 10-4 : <10-4 cpd/kg/kev (< 0.5% S m observed ) An effect from temperature can be excluded + Any possible modulation due to temperature would always fail some of the peculiarities of the signature Distribution of the relative variations of the operating T of the detectors

31 Summarizing on a hypothetical background modulation in DAMA/LIBRA 1-41 No Modulation above 6 kev A=(0.9±1.1) 10-3 cpd/kg/kev DAMA/LIBRA No modulation in the whole energy spectrum + if a modulation present in the whole energy spectrum at the level found in the lowest energy region R 90 tens cpd/kg 100 σ far away σ 1% No modulation in the 2-62 kev multiple-hits residual rate multiple-hits residual rate (green points) vs single-hit residual rate (red points) No background modulation (and cannot mimic the signature): all this accounts for the all possible sources of bckg Nevertheless, additional investigations performed...

32 Can a possible thermal neutron modulation account for the observed effect? Thermal neutrons flux measured at LNGS : Φ n = n cm -2 s -1 (N.Cim.A101(1989)959) Experimental upper limit on the thermal neutrons flux surviving the neutron shield in DAMA/LIBRA: studying triple coincidences able to give evidence for the possible presence of 24 Na from neutron activation: Φ n < n cm -2 s -1 (90%C.L.) 24m Na (T 1/2 =20ms) NO σ n = 0.43 barn σ n = 0.10 barn Two consistent upper limits on thermal neutron flux have been obtained with DAMA/NaI considering the same capture reactions and using different approaches. Evaluation of the expected effect: Capture rate = Φ n σ n N T < captures/day/kg HYPOTHESIS: assuming very cautiously a 10% thermal neutron modulation: S m (thermal n) < cpd/kg/kev (< 0.01% S m observed ) In all the cases of neutron captures ( 24 Na, 128 I,...) a possible thermal n modulation induces a variation in all the energy spectrum Already excluded also by R 90 analysis MC simulation of the process When Φ n = 10-6 n cm -2 s -1 : cpd/kg/kev cpd/kg/kev E (MeV)

33 Can a possible fast neutron modulation account for the observed effect? HYPOTHESIS: assuming - very cautiously - a 10% neutron modulation: NO In the estimate of the possible effect of the neutron background cautiously not included the 1m concrete moderator, which almost completely surrounds (mostly outside the barrack) the passive shield Measured fast neutron LNGS: Φ n = n cm -2 s -1 (Astropart.Phys.4 (1995)23) By MC: differential counting rate above 2 kev 10-3 cpd/kg/kev S m (fast n) < 10-4 cpd/kg/kev (< 0.5% S m observed ) Experimental upper limit on the fast neutrons flux surviving the neutron shield in DAMA/LIBRA: through the study of the inelastic reaction 23 Na(n,n ) 23 Na*(2076 kev) which produces two γ s in coincidence (1636 kev and 440 kev): Φ n < n cm -2 s -1 (90%C.L.) well compatible with the measured values at LNGS. This further excludes any presence of a fast neutron flux in DAMA/LIBRA significantly larger than the measured ones. Moreover, a possible fast n modulation would induce: a variation in all the energy spectrum (steady environmental fast neutrons always accompained by thermalized component) already excluded also by R 90 a modulation amplitude for multiple-hit events different from zero already excluded by the multiple-hit events Thus, a possible 5% neutron modulation (ICARUS TM03-01) cannot quantitatively contribute to the DAMA/NaI observed signal, even if the neutron flux would be assumed 100 times larger than measured by various authors over more than 15 LNGS

34 Radon Three-level system to exclude Radon from the detectors: Walls and floor of the inner installation sealed in Supronyl ( cm 2 /s permeability). Whole shield in plexiglas box maintained in HP Nitrogen atmosphere in slight overpressure with respect to environment Detectors in the inner Cu box in HP Nitrogen atmosphere in slight overpressure with respect to environment continuously since several years measured values at level of sensitivity of the used radonmeter Amplitudes for annual modulation of Radon external to the shield: <flux> 320 l/h Over pressure 3.1 mbar Time behaviours of the environmental radon in the installation (i.e. after the Supronyl), from which in addition the detectors are excluded by other two levels of sealing! DAMA/LIBRA-1 DAMA/LIBRA-2 DAMA/LIBRA-3 DAMA/LIBRA-4 Radon (Bq/m 3 ) -(0.029 ± 0.029) -(0.030 ± 0.027) (0.015 ± 0.029) -(0.052 ± 0.039) NO DM-like modulation amplitude in the time behaviour of external Radon (from which the detectors are excluded), of HP Nitrogen flux and of Cu box pressure Investigation in the HP Nitrogen atmosphere of the Cu-box Study of the double coincidences of γ s (609 & 1120 kev) from 214 Bi Radon daughter Rn concentration in Cu-box atmosphere < Bq/m 3 (90% C.L.) By MC: < low energy for single-hit events(enlarged matrix of detectors and better filling of Cu box with respect to DAMA/NaI) An hypothetical 10% modulation of possible Rn in Cu-box: < cpd/kg/kev (<0.01% S m observed ) An effect from Radon can be excluded + any possible modulation due to Radon would always fail some of the peculiarities of the signature and would affect also other energy regions

35 Can the µ modulation measured by MACRO account for the observed effect? Case of fast neutrons produced by muons Φ LNGS 20 µ m -2 d -1 Neutron LNGS: Y= n /µ /(g/cm 2 ) R n = (fast n by µ)/(time unit) = Φ µ Y M eff where: (±2% modulated) (hep-ex/ ) Annual modulation amplitude at low energy due to µ modulation: S m (µ) = R n g ε f E f single 2% /(M setup E) g = geometrical factor ε = detection efficiency by elastic scattering f E = energy window (E>2keV) efficiency f single = single hit efficiency S m (µ) < (0.4 3) 10-5 cpd/kg/kev NO Hyp.: Knowing that: M eff = 15 tons g ε f E f single 0.5 (cautiously) M setup 250 kg and E=4keV Moreover, this modulation also induces a variation in other parts of the energy spectrum It cannot mimic the signature: already excluded also by R 90

36 σ = 0.3% Noise Distribution of variations of total hardware rates of the crystals over the single ph.el. threshold (that is from noise to infinity ) during DAMA/LIBRA-1,2,3,4 running periods cumulative gaussian behaviour fully accounted by expected statistical spread arising from the sampling time used for the rate evaluation R Hj = hardware rate of j-th detector above single photoelectron <R Hj > = mean of R Hj in the corresponding annual cycle Amplitudes for annual modulation well compatible with zero: DAMA/LIBRA-1 DAMA/LIBRA-2 DAMA/LIBRA-3 DAMA/LIBRA-4 Hardware rate (Hz) -(0.20 ± 0.18) 10-2 (0.09 ± 0.17) (0.03 ± 0.20) 10-2 (0.15 ± 0.15) 10-2 Can a noise tail account for the observed modulation effect? Despite the good noise identification near energy threshold and the used very stringent acceptance window for scintillation events (this is only procedure applied to the data), the role of an hypothetical noise tail in the scintillation events has even been quantitatively investigated. The modulation amplitude of the "Hardware Rate" (period and phase as for DM particles) is compatible with zero: (0.03±0.09) 10-2 Hz < Hz (90% CL) Hardware Rate = noise +bckg [up to MeV]+signal [up to 6keV] noise/crystal 0.10 Hz relative modulation amplitude from noise < Hz/2.5 Hz (90%CL) even in the worst hypothetical case of 10% residual tail of noise in the data relative modulation amplitude from noise at low energy < NO <10-4 cpd/kg/kev

37 The calibration factors Distribution of the percentage variations (ε tdcal ) of each energy scale factor (tdcal k ) with respect to the value measured in the previous calibration (tdcal k-1 ) for the DAMA/LIBRA-1 to -4 annual cycles. Distribution of the percentage variations (ε HE ) of the high energy scale factor with respect to the mean values for the DAMA/LIBRA-1 to -4 annual cycles. ε tdcal tdcalk tdcal = tdcal k 1 DAMA/LIBRA-1,2,3,4 k 1 gaussian behaviours the low energy calibration factor for each detector is known with an uncertainty <<1% during the data taking periods: additional energy spread σ cal σ cal 1 σ cal / E 4 E σ = σ 1 ; res + σ cal σ res + σ res 2 σ res / E 20keV Negligible effect considering routine calibrations and energy resolution at low energy Confirmation from MC: maximum relative contribution < cpd/kg/kev σ=0.5% σ=0.6% No modulation in the energy scale + cannot mimic the signature Low-Energy calibration High-Energy factors (ε tdcal ) calibration factors (ε HE )

38 The efficiencies 2-8 kev σ = 0.3 % Distribution of variations of the efficiency values with respect to their mean values during DAMA/LIBRA running periods Time behaviour: modulation amplitudes obtained by fitting the time behaviours of the efficiencies including a WIMP-like cosine modulation for DAMA/LIBRA running periods Amplitudes ( 10-3 ) Energy DAMA/LIBRA-1 DAMA/LIBRA-2 DAMA/LIBRA-3 DAMA/LIBRA kev 4-6 kev 6-8 kev 8-10 kev (0.3±0.6) (0.0±0.6) -(0.3±0.6) -(0.5±0.5) (0.1±0.6) -(0.7±0.6) -(1.0±0.7) -(0.5±0.5) -(0.4±1.1) -(0.3±1.0) -(0.2±0.8) -(0.2±0.6) -(0.4±1.0) -(0.7±1.0) -(1.0±0.8) (0.7±0.6) Energy Modulation amplitudes (DAMA/LIBRA) 2-4 kev (0.1±0.4) kev -(0.4±0.4) 10-3 Amplitudes well compatible with zero + cannot mimic the signature

39 Summary of the results obtained in the additional investigations of possible systematics or side reactions (DAMA/LIBRA arxiv: ) Source Main comment Cautious upper limit (90%C.L.) RADON Sealed Cu box in HP Nitrogen atmosphere, < cpd/kg/kev 3-level of sealing, etc. TEMPERATURE Installation is air conditioned+ detectors in Cu housings directly in contact <10-4 cpd/kg/kev with multi-ton shield huge heat capacity + T continuously recorded NOISE Effective full noise rejection near threshold <10-4 cpd/kg/kev ENERGY SCALE Routine + instrinsic calibrations < cpd/kg/kev EFFICIENCIES Regularly measured by calibrations <10-4 cpd/kg/kev BACKGROUND No modulation above 6 kev; no modulation in the (2-6) kev <10-4 cpd/kg/kev multiple-hits events; this limit includes all possible sources of background SIDE REACTIONS Muon flux variation measured by MACRO < cpd/kg/kev + even if larger they cannot satisfy all the requirements of annual modulation signature Thus, they can not mimic the observed annual modulation effect

40 ... about the interpretation of the direct DM experimental results Presence of modulation for 11 annual cycles at ~8.2σ C.L. with the proper distinctive features of the signature; all the features satisfied by the data over 11 independent experiments of 1 year each one The positive and model independent result of DAMA/NaI + DAMA/LIBRA Absence of known sources of possible systematics and side processes able to quantitatively account for the observed effect and to contemporaneously satisfy the many peculiarities of the signature No other experiment whose result can be directly compared in model independent way is available so far To investigate the nature and coupling with ordinary matter of the possible DM candidate(s), effective energy and time correlation analysis of the events has to be performed within given model frameworks Corollary quests for candidates astrophysical models: ρ DM, velocity distribution and its parameters nuclear and particle Physics models experimental parameters a model or a model e.g. for WIMP class particles: SI, SD, mixed SI&SD, preferred inelastic, scaling laws on cross sections, form factors and related parameters, spin factors, halo models, etc. + different scenarios + multi-component halo? THUS uncertainties on models and comparisons

41 In progress complete model dependent analyses by applying maximum likelihood analysis in time and energy accounting for at least some of the many existing uncertainties in the field (as done by DAMA/NaI in Riv.N.Cim.26 n.1 (2003)1, IJMPD13(2004)2127, IJMPA21(2006)1445, EPJC47(2006)263, IJMPA22(2007)3155, EPJC53(2008)205, PRD77(2008)023506, arxiv: ), and to enlarge the investigations to other scenarios Just to offer some naive feeling on the complexity of the argument: experimental S m values vs expected behaviours for some DM candidates in few of the many possible astrophysical, nuclear and particle physics scenarios and parameters values

42 Examples for few of the many possible scenarios superimposed to the measured modulation amplitues S m,k 15 GeV N.F.W. 60 GeV N.F.W. WIMP DM candidate (as in [4]) considering elastic scattering on nuclei SI dominant coupling v 0 = 170 km/s DMp GeV Evans power law DMp N About the same C.L. scaling from NaI channeling contribution as in EPJC53(2008)205 considered for curve b [4] RNC 26 (2003) 1; [34] PRD66 (2002)

43 Examples for few of the many possible scenarios superimposed to the measured modulation amplitues S m,k 15 GeV N.F.W. 60 GeV N.F.W. WIMP DM candidate (as in [4]) Elastic scattering on nuclei SI & SD mixed coupling v 0 = 170 km/s DMp 100 GeV Evans power law DMp N About the same C.L. scaling from NaI θ = [4] RNC 26 (2003) 1; [34] PRD66 (2002)

44 Examples for few of the many possible scenarios superimposed to the measured modulation amplitues S m,k LDM candidate (as in arxiv: ): inelastic interaction with electron or nucleus targets m L =0 Light bosonic candidate (as in IJMPA21(2006)1445): a axion-like particles totally absorbed by target material About the same C.L. X-ray e - γ curve r: also pseudoscalar axion-like candidates (e.g. majoron) m a =3.2 kev g aee = [4] RNC 26 (2003) 1; [34] PRD66 (2002)

45 Conclusions DAMA/LIBRA over 4 annual cycles (0.53 ton yr) confirms the results of DAMA/NaI (0.29 ton yr) The cumulative confidence level for the model independent evidence for presence of DM particle in the galactic halo is 8.2 σ (total exposure 0.82 ton yr) The model-independent evidence by DAMA/NaI and DAMA/LIBRA well compatible with several candidates in several of the many possible astrophysical, nuclear and particle physics scenarios; other ones are open Available results from direct searches using different target materials and approaches and from searches for up-going µ do not give any robust conflict no model independent comparison Possible model dependent positive hints from indirect searches not in conflict with DAMA/NaI result Starting new data taking after upgrading to improve the investigation, to disentangle at least some of the many possibilities, to investigate other features of DM particle component(s) and second order effects, etc.. A possible highly radiopure NaI(Tl) multi-purpose set-up DAMA/1 ton (proposed by DAMA in 1996) is at present at R&D phase